10/1912 - Alex Hsu working on a virtual project through the Semiconductor Research fellowship
REVISED GOAL: 1/20/13
Figure out how to run a Virtual Screening Job on the AWS (Amazon Web Services) EC2 system using a linux virtual machine and VINA
Write a protocol for setting up the system
Write protocol for an end -user
Alex to get working on his free account - do a test dock of 30,000 ligands, create a Bestranking list and view top poses in PyMol.
Generate a step-by-step protocol for this
Then set it up in the VDS account and get it working there so that we can put the $400 amount from the URF (undergraduate research fellowship) there.
Amazon Web Services Notes (running notes for Alex and Dr. B to use):
Determine if Molecular Mechanics and/or Molecular Dynamics calculations improve the accuracy of ranking over regular virtual screening.
Specific Aims:
Find a good test system with which to validate (with positive and negative controls).
- use GOLD test set (on the GOLD website at CCDC) - however, these may not have 'wet lab' data to compare to (e.g. IC50s, or Ki values that correlate to true FEB energies).
- use other known binders to an example protein. Would need to have Ki, IC50 or deltaG info from ITC (Isothermal Titration Calorimetry)
- include list of 'dummy' atoms and/or negative controls that we don't expect to bind
Use GOLD, ICM, VINA to carry out regular docking
-- sa
Use AMBER on TACC to carry out calculations
Compare ranking lists of VS vs MM/MD calcs
Maybe do some wet lab verification of results
Maybe do a wet lab prospective study of novel ligands (do their wet lab binding affinities rank according to MM/MD calc rankings?)
Validation of Virtual Screening Methods
10/1912 - Alex Hsu working on a virtual project through the Semiconductor Research fellowship
REVISED GOAL: 1/20/13
Figure out how to run a Virtual Screening Job on the AWS (Amazon Web Services) EC2 system using a linux virtual machine and VINAWrite a protocol for setting up the system
Write protocol for an end -user
Alex to get working on his free account - do a test dock of 30,000 ligands, create a Bestranking list and view top poses in PyMol.
Generate a step-by-step protocol for this
Then set it up in the VDS account and get it working there so that we can put the $400 amount from the URF (undergraduate research fellowship) there.
Amazon Web Services Notes (running notes for Alex and Dr. B to use):
https://docs.google.com/document/d/1UT4D0rRKu1tKuAjjK9NorShrrSyOH-K9E1v56d7awZ4/edit?usp=sharingEC2 FAQ's:
http://aws.amazon.com/ec2/faqs/
Your AWS Account Alias is vdsclass
IAM users sign-in link:
https://vdsclass.signin.aws.amazon.com/console
GOAL:
Determine if Molecular Mechanics and/or Molecular Dynamics calculations improve the accuracy of ranking over regular virtual screening.Specific Aims:
Find a good test system with which to validate (with positive and negative controls).- use GOLD test set (on the GOLD website at CCDC) - however, these may not have 'wet lab' data to compare to (e.g. IC50s, or Ki values that correlate to true FEB energies).
- use other known binders to an example protein. Would need to have Ki, IC50 or deltaG info from ITC (Isothermal Titration Calorimetry)
- include list of 'dummy' atoms and/or negative controls that we don't expect to bind
Use GOLD, ICM, VINA to carry out regular docking
-- sa
Use AMBER on TACC to carry out calculations
Compare ranking lists of VS vs MM/MD calcs
Maybe do some wet lab verification of results
Maybe do a wet lab prospective study of novel ligands (do their wet lab binding affinities rank according to MM/MD calc rankings?)
REFERENCES & RESOURCES
AMBER site -http://ambermd.org/
CHARM site
This paper has a section on molecular modeling using the QM/MM method (from the Fast lab at UT Austin)
http://pubs.acs.org/doi/full/10.1021/bi201449r
Watowich/Malmstrom papers:
Denque research using TACC with Autodock and AMBER or CHARM for FEB (Free Energy of Binding) calculations